Epilepsy surgery involving the anteromedial temporal lobe (ATL) often results in verbal memory and naming decline following language dominant resection and in visual memory decline after nondominant resection. We have shown open ATL resections also frequently result in recognition or naming declines for several object types (e.g., famous persons) and related functions (e.g., proper noun fluency). These deficits, not previously recognized as clinical risks, compromise functional status and compound the burden of memory decline. We propose these deficits reflect collateral damage incurred when surgically accessing the mesial TL. Our ongoing NINDS K02 award uses DTI probabilistic tractography to determine if decline results from damage to specific white matter (WM) paths. This K02 is laying the groundwork for developing surgical options to improve cognitive outcome by sparing key paths underlying these functions. Preliminary data show several WM paths are related to baseline performance on these measures, and resection of specific tracts leads to predicted decline. Our R01 proposal uses a novel, minimally invasive, MRI-guided laser ablation technique, selective laser amygdalohippocampotomy (SLAH), to directly test our collateral damage hypothesis. SLAH ablates at least 50% of hippocampus and amygdala, but does not directly damage WM due to its minimally invasive nature. We hypothesize damage restricted to these structures will not lead to naming and recognition deficits, consistent with our preliminary data. This allows us to assess the popular view that mesial TL structures are essential to these processes, which has arisen from indirect paradigms (fMRI/stimulation mapping). Our clinical data also indicate most SLAH patients do not experience episodic memory decline, again contrary to prevailing views of hippocampal function. We propose better memory outcomes after SLAH results from: (a) sparing surrounding mesial TL structures; and (b) less direct and secondary degenerative damage to WM paths away from the resection zone. We predict patients undergoing SLAH will not experience naming and recognition deficits, and will exhibit a significantly better outcome than open resection patients. We predict WM tracts included in open resections will contribute to performance declines on these tasks, but not hippocampal and amygdalar resection volumes. Secondly, we evaluate the contribution of the hippocampus and amygdala to episodic memory functions, predicting greater memory deficits following open resection compared to SLAH. Lastly, we will explore secondary changes in non-resected WM tracts using MRI diffusion parameters, predicting greater improvement in these parameters after SLAH versus open resection. We hypothesize diffusivity improvements in these non-resected tracts and contralateral hippocampus will correlate with improved memory scores (greatest for the SLAH cohort). In sum, this research allows us to: (a) dissociate the contribution of amygdala and hippocampus from that of the overlying cortex and TL WM paths to key cognitive functions, and (b) determine if SLAH preserves cognitive functions frequently affected by open resections.